Multidimensional Monitoring of Recovery States, Biomarkers, Workload, and Performance in Elite Basketball: An Exploratory Longitudinal Field Study
Corazza G, Jamil S, Dema R, Zorba S, Bonetta L, Jòdar-Portas A et al. · Research Square · 2026
Abstract
Purpose: Elite team-sport programs increasingly combine athlete-reported wellness ratings with biochemical markers, yet the applied interpretation of these multimodal data remains uncertain. This exploratory longitudinal field study examined whether recovery-related athlete-reported states were associated with salivary markers, acute competition exposure, and performance-readiness profiles in elite basketball players.
Methods: The dataset comprised 309 observations from 13 elite basketball players monitored over 15 weeks. Athlete-reported variables included calmness, stress, anger, sleep quality, tiredness, muscle soreness, energy, and sleep duration. Salivary biomarkers included pH, uric acid, aspartate aminotransferase (AST), creatine kinase (CK), ketones, and immunoglobulin A (IgA). Minutes played were used as an indicator of acute competition exposure. Analyses included correlations and biomarker stratification, exploratory risk-enrichment analyses for biomarker abnormalities, fast-versus-slow subjective recovery comparisons, mixed-effects models of post-game-to-recovery change scores, and performance-readiness group comparisons.
Results: Severe muscle soreness and low energy were the clearest subjective flags of biomarker abnormality. The strongest enrichment was observed for low IgA among severe-soreness observations, with a rate of 75.0% compared with 14.9% among trait-negative observations. Salivary pH differentiated fast and slow recovery from tiredness (p = 0.0167), whereas AST differentiated recovery phenotypes for energy (p = 0.0130), calmness (p = 0.0240), and anger (p = 0.0385). In mixed-effects models adjusted for minutes played, tiredness change was negatively associated with salivary pH, while anger change remained positively associated with ketones and CK. Selected associations involving tiredness and calmness also varied according to minutes played. Good- and poor-performance-readiness observations showed significantly different multivariate profiles (Wilks lambda = 0.3564, F = 25.02, p